Population-based analysis of pathological correlates of dementia in the oldest old

Transcription

1 RESEARCH ARTICLE Population-based analysis of pathological correlates of dementia in the oldest old Maarit Tanskanen, Mira M akel a, Irma-Leena Notkola, Liisa Myllykangas, Sari Rastas, Minna Oinas, Perttu J. Lindsberg 5,6, Tuomo Polvikoski 7, Pentti J. Tienari 5,6 & Anders Paetau Department of Pathology, Huslab, Helsinki University Hospital and Medicum, University of Helsinki, Helsinki, Finland Finnish Information Centre for Register Research, Helsinki, Finland Mehil ainen Corporation, Helsinki, Finland Neurosurgery, Clinical Neurosciences, Helsinki University Hospital and University of Helsinki, Helsinki, Finland 5 Neurology, Clinical Neurosciences, Helsinki University Hospital and University of Helsinki, Helsinki, Finland 6 Biomedicum Helsinki, University of Helsinki, Helsinki, Finland 7 Institute of Neuroscience, Newcastle University, Newcastle upon Tyne, United Kingdom Correspondence Maarit Tanskanen, Department of Pathology, Huslab, Helsinki University Hospital and Medicum, University of Helsinki, Haartmaninkatu, PO Box, 000 University of Helsinki, Helsinki, Finland. Tel: ; Fax: ; maarit.tanskanen@hus.fi Funding Information This study was supported by Finska l akares allskapet, Liv och H alsa Foundation, P aivikki and Sakari Sohlberg Foundation and Helsinki University Central Hospital competitive research fund (EVO). Received: 6 May 06; Revised: November 06; Accepted: 6 December 06 Annals of Clinical and Translational Neurology 07; (): 5 65 doi: 0.00/acn.89 Abstract Objective: The aim of this study was to analyze brain pathologies which cause dementia in the oldest old population. Methods: All 60 persons aged 85 years living in the city of Vantaa (Finland), on April st, 99 formed the study population of the Vantaa85 + study, 00 of whom were autopsied during follow-up (79.5% females, mean age-at-death 9.7 years). Alzheimer s disease (AD) pathology (tau and beta-amyloid [Ab]), cerebral amyloid angiopathy (CAA) and Lewy-related pathologies were analyzed. Brain infarcts were categorized by size (< mm, 5 mm, >5 mm) and by location. Brain hemorrhages were classified as microscopic (< mm) and macroscopic. Results: 95/00 (65%) were demented. 9/95 (99%) of the demented had at least one neuropathology. Three independent contributors to dementia were identified: AD-type tau-pathology (Braak stage V-VI), neocortical Lewy-related pathology, and cortical anterior 5 mm infarcts. These were found in %, %, and % of the demented, respectively, with the multivariate odds ratios (OR) for dementia 5.5,.5, and.. Factor analysis investigating the relationships between different pathologies identified three separate factors: () AD-spectrum, which included neurofibrillary tau, Ab plaque, and neocortical Lewy-related pathologies and CAA () > mm cortical and subcortical infarcts, and () < mm cortical microinfarcts and microhemorrhages. Multipathology was common and increased the risk of dementia significantly. Interpretation: These results indicate that AD-type neurodegenerative processes play the most prominent role in twilight cognitive decline. The high prevalence of both neurodegenerative and vascular pathologies indicates that multiple preventive and therapeutic approaches are needed to protect the brains of the oldest old. Introduction People aged 85 years old represent the fastest growing age group in Western countries, projected to increase.5- fold in the US by 050 (The U.S. Census Bureau 0, Population.aspx). Because of the consequences of aging such as dementia these people are likely to need help of the health and social care services. Care of people with dementia is a major economic burden in Western societies. Its total cost in the US was 60 billion US$ in 00. Therefore, prevention of dementia should be one of the main aims of the health care service. The most common dementing disorder is Alzheimer s disease (AD), with estimated prevalence of 5% among the US citizens aged 85 in 0. Vascular dementia (VaD) and dementia with Lewy bodies (DLB) were estimated to comprise about one-third of all new dementia 5 ª 07 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

2 M. Tanskanen et al. Pathological Correlates Of Dementia cases in the US. However, various pathologies may coexist in the brains of elderly individuals. For example, twothirds of elderly AD patients have concomitant other brain pathologies such as cerebrovascular disease or Lewy-bodyrelated,5,6 potentially resulting in mixed dementia. 7 Previous neuropathological studies of mixed dementia have mostly focused on hospital-based series on 85 years old. To develop preventive measures against major components of dementia one must determine in which relation they prevail and coexist in the general old population, when clinically diagnosed dementia is present or absent. Recently, community-based studies 8,9,0 have added knowledge on the effect of multiple pathologies on dementia in the very old. We have addressed this question by performing a population-based and prospective neuropathological study on very elderly subjects (Vantaa 85+ study), to analyze the frequency and co-occurrence of common old-age-associated neuropathologies and their impact on the development of clinical dementia. Materials and Methods Vantaa 85+ study protocol and participants Vantaa 85+ Study included all individuals aged 85 years living in the city of Vantaa in Southern Finland on April, 99. The clinical follow-up studies were carried out in 99, 996 and 999 for the survivors. Of the 60 eligible subjects, 55 participated at least in the first clinical followup study. Postmortem neuropathological examinations were performed in 0/60 (50.9%). Four were excluded due to the insufficient clinical data, resulting in 00 subjects comprising the autopsied subpopulation of this study (Table S). Clinical examination The diagnosis of dementia followed the criteria of the Diagnostic and Statistical Manual of Mental Disorders, th edition, revised (DSM III-R; American Psychiatric Association, 987). The diagnosis was based on the Mini-Mental State Examination (MMSE), required the consensus opinion by two neurologists and had to be set earlier than months before death. Median time between the diagnosis of dementia and death was 6 months, SD months. There were demented subjects, of whom 95 were autopsied. Apolipoprotein E (APOE) genotyping was carried out as previously described. Macroscopic examination of the brain The brains were fixed for at least weeks in % phosphate-buffered formaldehyde. After external examination, the cerebrum, brain stem, and cerebellum were separated from each other. The cerebrum was cut in -cm-thick coronal and the brain stem and cerebellum in 5-mmthick sagittal slices. The presence, location, and number of macroscopic small ( 5 mm) or large (>5 mm) infarcts in the cerebral cortex, white matter, basal ganglia region, cerebellum and brainstem were categorized as hemispheric and deep subcortical. Microscopic analyses Neurodegenerative lesions: Alzheimer s disease and Lewy-related pathologies The analysis of AD-related pathologies was performed following the original criteria.,5 Density of the neuritic senile plaques (CERAD score) in sections stained with modifed Bielschowsky silver method was scored according to the Consortium to Establish Registry for Alzheimer s Disease (CERAD) criteria. Neuritic senile plaques were detected as silver-positive neocortical aggregations which may or may not have surrounded an intensively staining central core, and also included at least one intensively staining elongated structure orientating from the periphery toward the centre of the aggregate. Braak stages for the AD-related neurofibrillary tangles were defined using Gallyas silver-stained 80-lm-thick sections cut from polyethylene glycol-embedded samples from the entorhinal cortex at the level of the mammillary bodies, from the hippocampus at the level of the lateral geniculate body, and from the occipital lobe, so that the striate area, parastriatale field, and peristriate region were all represented in the sample. 5 The quantities of the plaques and tangles were counted in four specimens (middle frontal, superior temporal, and middle temporal gyri and inferior parietal lobule). The amount of b-amyloid protein was analyzed by quantifying the fraction of the cortical specimens covered by methenamine silver-stained plaques. Contiguous cortical microscopical fields were examined, extending from the pial surface to the white matter, at a magnification of 00 with a square microscopical graticule. 7 The tangles were counted in the modifed Bielschowsky silverstained sections in contiguous cortical microscopical fields, extending from the pial surface to the white matter, at a magnification of 00 with a square microscopical counting frame, 0.55 mm in width, along a line perpendicular to the pial surface. 7 The Lewy-related pathology was assessed according to the consensus criteria 8 using histological staining methods and IHC against a-synuclein clone. 9 Vascular lesions: cerebral amyloid angiopathy, microscopic ischemic lesions, and hemorrhages The diagnosis of cerebral amyloid angiopathy (CAA, the deposition of amyloid b (Ab) -derived amyloid in the ª 07 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. 55

3 Pathological Correlates Of Dementia M. Tanskanen et al. subarachnoid and intracortical small and middle-sized blood vessels) in six brain regions (frontal, parietal, temporal and occipital lobes, hippocampus, and cerebellum) was based on Congo red, confirmed using IHC against Ab peptide clone G8. 0 The severity of CAA was assessed by counting the percentage of the Congo Red-positive and Ab immune reactive small and middle-sized vessel profiles of all vessel profiles seen in samples obtained from each region. The counts were combined to create the median value for all six regions. The values higher than the upper quartile (Q) value (5.% for all six brain regions and 6% for the frontal lobe) were defined as severe forms of CAA. The presence and count of old cortical microinfarcts and microhemorrhages were investigated in six brain regions (above). The microinfarcts, defined as focal lesions mm and invisible to the naked eye showing neuronal loss, glia cell, and macrophage reaction and/or cystic tissue necrosis, were evaluated in the Hematoxylin Eosin stained 6-lm-thick tissue sections. The presence of iron containing macrophages, reflecting the microhemorrhage, was verified using the Prussian blue stain. These microscopic diagnoses were set by at least two pathologists (AP, MT, or MM). Statistical analyses Statistical analyses were performed using IBM SPSS Statistics version 0 (IBM, New York, USA) software. Logistic regression was used to study the associations between different neuropathologies and dementia (Tables,). Reciprocal correlations between the neuropathological and genetic variables were analyzed using Spearman bivariate correlation analysis (Table ). The variance in eight quantitative variables was analyzed using factor analysis with the principal component analysis and rotation method (Fig., Table S). Statistical analyses were performed by two investigators (MT, I-LN). Approval for the study The Vantaa 85+ study was approved by the Ethics Committee of the Health Centre of the City of Vantaa. The Finnish Health and Social Ministry approved the use of the health and social work records and death certificates. The National Authority for Medicolegal Affairs (VALVIRA) has approved the collection of the tissue samples at autopsy and their use for research. Blood samples were collected after having obtained an informed consent from the subjects or their relatives. The consent for autopsy was obtained from the nearest relative. Results Prevalence of different types of neuropathologies in the population 9/95 of the demented and 05/05 of the nondemented subjects had at least one type of neuropathology (any severity, any location) of the five categories shown in Table : () AD-type () Lewy-related, () brain infarct, () CAA, and (5) brain hemorrhage. Thus, with the exception for one demented subject, all 00 subjects had at least one type of neuropathology. The prevalence of each neuropathology type/subtype in the demented and nondemented subjects is presented in Table. Neuropathological variables and dementia Dementia was diagnosed in 95/00 (65%). The associations between individual neuropathological categories and dementia were first tested with univariate analysis (Table ). Among the evaluated parenchymal AD-type pathologies, Braak stage V VI had the strongest association with dementia (OR: 0.76, 95% CI:. 7.5). Lewyrelated pathology, when spread to the neocortex had also a strong and significant association with dementia (OR:.78; 95% CI:.9.78), whereas Lewy-related pathology in the brainstem and limbic regions was not associated with dementia. Of the brain infarcts, 5 mm cortical infarcts (highest OR:.8, 95%, CI:. 6.0 for cortical infarcts in the estimated territory of the anterior circulation) or thalamic infarcts (OR:.5, 95% CI: ) associated with dementia. CAA in any brain region was associated with dementia (OR:.7, 95% CI:.7.77) but the association was strongest for severe CAA in the frontal lobe (OR:.67, 95% CI: ). CERAD scores C and B were also associated with dementia in the univariate analysis (OR: 7.8 (-75-.) and.78 (.56.96). In order to identify the independent contributions of each pathology to the prediction of dementia, we applied a multivariate analysis (Table ). In the multivariate model the variable with the strongest association with dementia from each pathology category was selected as a covariate and these were tested together in the same analysis. From the AD-type pathology group, however, we included both Braak stage and CERAD score, both as categorical variables (Table ), and in total there were five strong pathologies that were analyzed in multivariate analysis, in addition to gender, age at death, and the carriership of the APOE e. Three of these five pathological variables associated independently with dementia in the multivariate analysis (Table ): () Braak stage V VI, () diffuse neocortical type of Lewy-related pathology, and 56 ª 07 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association.

4 M. Tanskanen et al. Pathological Correlates Of Dementia Table. Frequency of various neuropathologies and association of such pathologies with dementia in the autopsied subpopulation in the Vantaa 85 + study. Autopsied population N = 00 Dementia Neuropathology present Demented N = 95 Nondemented N = 05 OR (95% CI) P -value Univariate OR (95% CI) P -value Multivariate Alzheimer type Braak stage Braak stage 0 II 5 (.) (.9) Reference Braak stage III IV 8 (.) 55 (5.).9 ( ) 0..6 ( ) Braak stage V VI 66 (.8) 6 (5.7) (.0 7.5) < ( ) CERAD score CERAD score 0 (7.0) 8 (6.) Reference CERAD score A 7 (8.7) 6 (5.). ( ) ( ) 0.9 CERAD score B (56.9) 6 (.8).779 ( ) ( ) 0.8 CERAD score C (7.) 5 (.8) 7.80 (.75.7) < ( ) 0.70 Lewy-body-related: DLB category 5 Lewy bodies absent (brainstem) 0 (6.6) 8 (80.0) Reference NI Lewy bodies present Brainstem predominant/limbic type (7.) 5 (.) NS NI Diffuse neocortical type (.0) 6 (5.7).78 (.9.78) ( ) Infarct 6 Absent 7 (6.) 9 (6.7) Reference NI Present (6.6) 56 (5.) NS NI Superficial total 85 (.6) (0.5).76 (.07.9) 0.07 NI Superficial > 5 mm 0 (0.5) 5 (.) NS NI Superficial 5 mm 65 (.) (.0).87 (.08.9) 0.05 NI Superficial cortical 5 mm 5 (.) (0.5).56 (.6 5.) NI Superficial cortical anterior 5 mm (.0) 9 (8.6).8 (. 6.0) (.68 9.) 0.05 Superficial cortical posterior 5 mm (5.9) 8 (7.6).9 (.0 5.9) 0.07 NI Deep subcortical total 76 (9.0) (0.5) NS NI Deep subcortical >5 mm 0 (5.) 6 (5.7) NS Deep subcortical 5 mm 7 (6.9) 9 (7.6) NS NI Thalamic 5 mm 9 (.9) 7 (6.7).5 ( ) 0.0 NI Microscopic cortical < mm (7.) 6 (5.) NS NI Cerebral amyloid angiopathy (CAA) Absent in all 6 brain regions 6 7 (.) (.9) Reference NI Present In any 6 7 brain regions 8 (75.9) 6 (58.).7 (.7.77) 0.00 NI In the frontal lobe (6.6) 5 (.9). (..78) 0.00 NI Severe CAA Severe CAA in 6 brain regions 8 60 (0.8) 5 (.).67 (..98) 0.00 NI Severe CAA in the frontal lobe 9 6 (.8) (.).6 ( ) < ( ) Cerebral hemorrhage Absent 78 (0.0) (.) Reference Present 7 (60.0) 7 (67.7) NS NI Macroscopic (.) (.9) NS NI Cortical microscopic 5 (59) 70 (66.7) NS NI Cortical microscopic with >5 siderophages 0 (5.) 6 (5.7) NS NI (Continued) () cortical 5 mm infarcts in the anterior parts of the brains. Braak stage V VI was the most significant risk factor for dementia both in the univariate and multivariate analyses (Table ). Severe frontal CAA showed a borderline association with dementia in the multivariate analysis (P = 0.057), whereas CERAD score and APOE e-carrier status were not independently associated with dementia in the multivariate model (P < 0.). ª 07 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. 57

5 Pathological Correlates Of Dementia M. Tanskanen et al. Table. Continued. Autopsied population N = 00 Dementia Neuropathology present Demented N = 95 Nondemented N = 05 OR (95% CI) P -value Univariate OR (95% CI) P -value Multivariate Covariates Age at death, mean (SD) 9.5 (.7) 9. (.8) NS NS Male gender 0 (5.) (.0) NS NS Apolipoprotein E e carrier 7 (9.6) 6 (6.7).7, (<0.00).096 ( ) 0.86 OR, Odd s ratio; CI, confidence interval; NI, not included; NS, nonsignificant; DLB, dementia with Lewy bodies. Bivariate logistic regression analysis. Within each neuropathology category (Alzheimer, Lewy body, etc.) only the variable with strongest association with dementia was selected in the multivariate model. However, for AD-type pathology, both Braak stage and CERAD score were included and analyzed as categorized variables. Other included variables were diffuse neocortical type of the Lewy body -related pathology, cortical anterior infarcts 5 mm, severe CAA in the frontal lobe 0. These selected variables were analyzed together with other covariates (age at death, gender, and the possession of Apolipoprotein E e allele). Subjects having neurofibrillary pathology in the neocortex versus subjects with such pathology in hippocampus, entorrhinal cortex, or limbic area or without such pathology (Braak 5 ). Subjects with frequent plaques in the neocortex versus subjects with no to moderate plaques (Mirra et al., ). 5 McKeith et al., 8 6 After: Troncoso et al. We used the designation superficial for cortical and superficial subcortical lesions and the designation deep subcortical for lesions situated in the basal ganglia, brainstem, cerebellum, and capsula interna. 7 Frontal, temporal, parietal and occipital lobe, hippocampus, cerebellum. 8 Congo red positivity and Ab immunoreactivity detected in 5.% (Q value) of all vessel profiles in the 6six brain regions (Tanskanen et al., 0 ). 9 Congo red positivity and Ab immunoreactivity detected in >6% (Q value) of the vessel profiles in the frontal lobe (Tanskanen et al., 0 ). Table. The effect of combined counts of four most significant neuropathologies on dementia. Autopsied population N = 00 Number of pathologies Demented N = 95 Prevalence of multipathology Nondemented N = 05 OR (95% CI), 0 65 (.) 80 (76.) Reference 69 (5.) 8 (8.) 5. ( ) (.6) 6 (7.0) 9.9 (. 0.9) 5 (7.7) (.) 0.8 (.6 9.0) (.0) 0 (0.0) n.a. 6 (.) 7 (8.0) 0.07 ( ) OR, Odd s ratio; CI, confidence interval; n.a., not applicable. Significant. Pathologies associating significantly with dementia (Braak stages V VI, the presence of diffuse neocortical type of Lewy-body-related pathology, frontal CAA > Q value (6%), and macroscopic infarcts 5 mm in the cortex supplied from anterior and middle cerebral arteries, see: Table ). Bivariate logistic regression analysis. Adjusted for age at death, gender and the possession of Apolipoprotein E e allele. We next analyzed the co-occurrence of different types of neuropathologies. We included in the analysis the three independent neuropathological contributors of dementia (Braak stage V VI, diffuse neocortical type of Lewyrelated pathology, and anterior cortical 5 mm infarcts). In addition, as a representative of the Ab pathology, severe frontal CAA was included, because the association between dementia and severe frontal CAA in the multivariate analysis was nearly significant, whereas the association between CERAD score and dementia was insignificant (Table ). One-third of the demented had at least two and two-thirds had at least one of these most significant variables (Table ). The risk for dementia of subjects with two or more of these four pathologies was almost two times higher than in those with only one of these pathologies (OR: 0.07, 95% CI: vs % CI: ). On the other hand, one-third of the demented subjects did not have any of these pathologies. The distribution of these pathologies at an individual level is shown in Figure, highlighting that the distribution of these neuropathologies is random. The combination of tau- and neocortical Lewy-body-related pathologies produced the highest risk (Table S). Relationships between the various neuropathologies Correlation analysis (Table ) revealed correlations between the AD-type pathologies, neocortical Lewyrelated pathology and CAA. Different types of brain 58 ª 07 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association.

6 M. Tanskanen et al. Pathological Correlates Of Dementia Table. Results of correlation analyses between the various quantitative neuropathological variables and the possession of apolipoprotein E e allele in the whole autopsied subpopulation (N = 00) of the Vantaa 85 + Study. Plaque 5 6 Deep sc 8 < mm 9 0 AD type Lewy rel Infarct (N) CAA Hem (N) APOE Tangle Neocortical Hemisph 7 6 regions cortical e-carrier Microsc. Microsc. AD type Tangle Plaque Lewy rel Neocortical Infarct (N) Hemisph Deep sc Microsc < mm 9 CAA 6 regions Hem (N) Microsc. cortical APOE e carrier The counts represent correlation coefficients (Spearman s q). AD, Alzheimer s disease; Lewy rel, Lewy-related pathology; CAA, cerebral amyloid angiopathy; Hem, hemorrhage; APOE, Apolipoprotein E; Tangle, neurofibrillary tangle, Plaque, methenamine-silver-positive plaque (any type); Hemisph, hemispheric; Deep sc, deep subcortical; Microsc, microscopic. Significant at the level. Significant at the 0.00 level. Percentage of the amyloid laden blood vessels of all vessels (Tanskanen et al., NAN 0; 8: 9-6). Average number of neurofibrillary tangles in the four cortical regions (frontal, temporal, parietal, occipital). 5 Percentage of neocortex covered by methenamine silver-positive plaques (frontal, temporal, parietal, occipital, Polvikoski et al., 6 ). 6 The presence or absence of neocortical Lewy bodies in the temporal, frontal and parietal lobes and cingulate gyrus (McKeith et al., 8 ). 7 Macroscopic infarcts in the cerebral cortex and superficial subcortical white matter (Troncoso et al., ). 8 Macroscopic infarcts in the basal ganglia, brainstem, and cerebellum (Troncoso et al., ). 9 Microscopic cortical infarcts < mm. 0 Frontal, temporal, parietal, occipital lobe, hippocampus, cerebellum. infarcts correlated with each other and APOE e correlated with AD-type parenchymal pathology, CAA and microscopic (< mm) cortical infarcts. To investigate the relationships of different pathologies further we used factor analysis. This analysis identified three separate factors: () AD-spectrum, which included neurofibrillary tau pathology, Ab plaque pathology, CAA, and neocortical Lewy-related pathology, () > mm cortical and subcortical infarcts, and () < mm cortical microinfarcts and microhemorrhages (Fig., Table S). Discussion We have performed a study on the spectrum of various neuropathologies underlying dementia in a populationbased study on subjects aged 85 years. Our major findings were: () the prevalence of any age-associated brain pathology was extremely high (99.7%), () neurofibrillary AD-type pathology (Braak stage V VI) was the strongest independent neuropathological correlate of dementia, although two other pathologies (neocortical type of Lewyrelated pathology, and 5 mm cortical infarcts especially in the territory of the anterior circulation) also associated independently with dementia, () the different types of neuropathologies formed up into three separate, frequency-based clusters. () multipathology is common in the very elderly demented individuals and increases the risk of dementia significantly. Neuropathological and clinical diagnoses Autopsy with neuropathological examination is often needed to determine the background of dementia, in addition to the clinical and radiological examinations. Today, neuropathological examination is a prerequisite for the definite diagnosis of AD. However, in the very elderly, the differential diagnostics of dementia syndromes may be challenging. 7 Although the contemporary neuropathological diagnosis of AD notifies some other pathologies than the AD-related, such as DLB with consensus neuropathological criteria 8 there are conditions such as VaD that lack such criteria. Therefore, we used the clinical dementia diagnosis as a starting point, and tested the presence versus not presence of clinical dementia against the panel of neuropathological components. The criteria for the clinical diagnosis of dementia can be considered strict enough to reliably differentiate between the demented and nondemented. Our neuropathological ª 07 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. 59

7 Pathological Correlates Of Dementia M. Tanskanen et al. Factor Factor Plot in Rotated Factor Space,0 0,5 0,0-0,5 -,0 Hemispheric infarcts -,0-0,5 0,0 Factor MH MI 0,5,0 Deep subcortical infarcts Lewy bodies Tangles CAA Plaques,0 0,5 0,0-0,5 -,0 Factor Figure. The variance in eight quantitative variables, as analyzed using factor analysis: The variables (components) investigated consisted of counts of neurofibrillary tangles (tangles), percentage of cortical thickness covered by methenamine silver-positive senile plaques (plaques), hemispheric and deep macroscopic infarcts, cortical micro infarcts < mm (MI), cortical microhemorrhages (MH), and the severity of cerebral amyloid angiopathy (CAA) in six brain regions, and the presence/absence of neocortical Lewy bodies, as illustrated in a three-dimensional form. Three factors explaining the variability between the eight components investigated were extracted (Table ) showing that the cluster of four variables (Factor, red color), consisted of tangles, plaques, CAA and the neocortical type of Lewybody-related pathology, explained the majority of the variability between the variables. analyzes followed the traditional neuropathological parameters for AD,6 and LBD. These criteria have recently been refined and adjusted for the needs of clinical diagnostics, for example, by applying IHC-based methods but the original methods can still be judged reliable indicators of the severity of the pathology.,, For the assessment of CAA several methods have been proposed. Here we have preferred the above described, allowing quantitative analyses based on whole tissue sections. 0 The neuropathological spectrum of the aged brain The extremely high prevalence of any neuropathology (99.7%) shows that the development of at least some type of neuropathology during aging seems inevitable. The extremely high frequency of the AD-related pathologies (9/00 [97%] with Braak stage >0 or CERAD score>0, strongly contributed to this count similarly as in another population-based study (Honolulu Asia Aging Study, HAAS) on elderly subjects. 5 Other population-based studies had found somewhat lower frequencies of ADtype pathologies,,6 8, probably as those studies included younger subjects, with lower mean age at death, as compared to our study (9. years for the autopsied population, Table S). Indeed, Strotzyk et al. reported AD in 55/90 of subjects in a community-based series with mean age at death was 8. years. 9 Other common pathologies were CAA (any severity 09/00 [70%]), cortical microhemorrhages (85/00 [6%]), and brain infarcts (80/00 [60%]). The significance of CAA in dementia has been somewhat unclear. Previous population-based studies,0, have demonstrated higher prevalence of CAA in the demented as compared with the nondemented, referring to an impact for CAA for the development of dementia. However, only one populationbased study has demonstrated an independent association between CAA and dementia. 6 This discrepancy 8 may reflect the effect of variability in the selection of variables in the multivariate analyses. Some brain lesions, such as cortical microhemorrhages, although associated with age did not associate with dementia. There are only few histological studies on MH and dementia. One previous study found microhemorrhages to be frequent in AD brains in contrast to us (Table ). Hemorrhage, also microhemorrhage, can be a manifestation of CAA but here we did not find a statistical association between the presence of microhemorrhage and CAA (Table ). One reason for these discrepancies may be that the presence of microhemorrhage was analyzed here in the cortical regions only. The significant neuropathologies In this study, the AD-related tau-pathology (Braak stages V VI, Table ) was the strongest single contributor to dementia, but the impact of the neocortical Lewy-related pathology was also strong. These findings are consistent with previous longitudinal studies investigating neuropathological correlates of dementia.,6,7,8,9,5,6,7 Several studies also have demonstrated a strong association between high CERAD score and dementia, but the results in the multivariate analyses have varied. Similar to our study, the association was lost in one study 7, in contrast to another 6 in which the strong association prevailed in the multivariate analysis. These discrepancies may reflect the differences in the variables selected in the multivariate analyses. We confirmed the previous findings on the independent association between dementia and cerebral small macroscopic infarcts ( 5 mm),6, emphasizing the role of vascular backgrounds in dementia in the very elderly. Indeed, not only cerebral infarcts but even 60 ª 07 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association.

8 M. Tanskanen et al. Pathological Correlates Of Dementia atherosclerosis of the cerebral large and small vessels have recently shown to increase the risk of AD-type dementia. 8 However, the association between dementia and the presence of 5 mm cortical infarcts in the anterior parts of the brain was even stronger (OR:.8, 95% CI:. 6.0), and represented an independent determinant of dementia, thus referring to the role of lesion localization in the clinical outcome, independent from the background (an obstruction/decreased blood flow in either the local small blood vessels or larger arteries). The association between cerebral microscopic infarcts and dementia, found in several previous studies,,7,6 was not confirmed here, most probably as we have not analyzed white matter lesion but cortical microscopic infarcts only. Although deep infarcts as a group did not associate with dementia, similar to another longitudinal study, we showed the significance of infarcts in the thalamic region in cognitive impairment (Table ). The network of various neuropathologies Factor analysis has mainly been applied in the social sciences, for instance, in the opinion polls. It measures the conceptual distance between different quantitative variables relative to each other. Here we applied the method for the first time on a neuropathological study, to identify the settling of the quantitative variables in relation to the others to form variable clusters. The resulting perspective (Fig., Table S) recapitulated the results of the correlation analysis (Table ). The AD-type pathology, CAA, and Lewy-related pathologies were separated into their own cluster (Factor, AD-spectrum ). The placing of the Lewy-related pathology in the AD-spectrum is not unexpected, as the Lewy-related pathology when present tends to coexist with AD pathology. 9 This also explains the high association between the combinations of these pathologies and dementia. The natural background for the second cluster (Factor, the macroscopic > mm infarcts) is arteriosclerosis/thrombosis of the arterioles and larger arteries. The third cluster (Factor ) consisted of cortical microinfarcts and microhemorrhages. Although both components of this cluster potentially represent a type of capillary injury neither was associated with dementia. Thus, apart from CAA, the various vascular pathologies in this study did not seem to be linked with the AD- or Lewy-related pathologies, suggesting that these vascular and primary neurodegenerative pathologies mainly are independent from each other, as suggested previously. 5 Multipathology We show the extent (Tables,) and spectrum (Fig., and Table S, S) of the concomitant pathologies in the brains of the oldest old, demonstrating the high frequency of multiple pathologies, as has been reported before. 7 The increase in the number of pathologies increased the OR for dementia (Table, Figure ), similarly to other recent community-based studies. 8,9,0 It is of interest that the effect on multipathology on dementia was similar in spite of differences in the pathologies included in the mixed pathology group in the different studies. For example, in the study of Kawas et. al. 9, the subjects were categorized in four different groups, based on the presence and severity of AD pathology and the presence of other type of pathologies, whereas another study 0 focused on the presence of AD and TDP- pathologies. These results imply that the association between multipathology and dementia is not restricted to specific neuropathologies or their combinations but several different combinations of several different neuropathologies may increase the risk of dementia. Second, the clinical diagnosis and especially the determination of the type of dementia in the very old may be challenging, 0 as many patients fulfill more than one diagnostic criterion and more than one pathophysiology. Third, it is possible that the co-occurrence of multiple pathological processes in the brain and their additive effect in the risk of dementia may also have a deteriorating effect on the reliability of short term drug trials lacking neuropathological confirmation of the diagnoses. 7 On the other hand, probably as a result of this multipathology, many modifiable risk factors for dementia have been identified and several of them are () shared with atherosclerotic cerebrovascular disease and () amenable to simple preventive measures of, for example, lifestyle and nutrition. Even though we found that two-thirds of the demented subjects had at least one of the four pathologies that associated independently or nearly independently with dementia (Braak stage V VI, diffuse neocortical type of Lewy-related pathology, anterior cortical 5 mm infarcts and frontal severe CAA), it is noteworthy that one-third of the demented did not have any of these pathologies. On the other hand, 99% of the demented had at least one neuropathology analyzed here, of which some, for example, the presence of any 5 mm cortical infarcts or CERAD score C may explain dementia in these individuals. It is also of note that our study lacked data on some important dementia-associated neuropathologies, for example, hippocampal sclerosis and the other TDP--related pathologies as well as primary age-related tauopathy (PART), the analysis of which are not completed yet in our study material. Further studies, including these and other, for example, vascular neuropathologies may further specify the neuropathological backgrounds of dementia in the very elderly. ª 07 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. 6

9 Pathological Correlates Of Dementia M. Tanskanen et al. Demented, Braak V-VI DemFrcaa Severe frontal CAA Nondemented, Braak V-VI 0 N = 66 DemAnCo Small corti- Incal infarcts DemNASY Neocortical Lewy bodies DemBr Braak V-VI 0 N = 6 SeFrCA Severe frontal CAA Nond NeASNo Small cortical infarcts nd BrNonde Braak V-VI 0 Demented with neocortical Lewy bodies DemBr Braak V-VI Severe frontal CAA DemFrca a Small cortical infarcts DemAnC oin DemNAS Neocortical Lewy bodies Y N = 0 Nondemented with neocortical Lewy bodies N = 6 SeFrCA Severe frontal Nond CAA NeASNo Neocortical ndlewy bodies 0 Demented with small cortical infarcts DemNASY Neocortical Lewy bodies DemBr Braak V-VI DemFrcaa Severe frontal CAA DemAnCoI Small corti- infarcts 0 ncal N = Nondemented with small cortical infarcts N = 9 BrNonde Braak V-VI Severe SeFrCANo fron-tal ndcaa AnCoInN Small cortiondem cal infarcts 0 Demented, severe frontal CAA N = 6 DemAnCoIn Small cortical infarcts DemNASY Neocortical Lewy bodies DemBr Braak V-VI DemFrcaa Severe frontal CAA 0 Nondemented, severe frontal CAA N = Small cortical infarcts m NeASNond Neocortical Lewy bodies BrNonde Braak V-VI SeFrCANond Severe frontal CAA Figure. Distribution of four dementia-associated pathologies in demented and nondemented: Distribution and co-occurrence of the three neuropathologies which associate independently with dementia: Braak stages V-VI (blue color), neocortical Lewy-related pathology (green color), and the presence of 5 mm infarcts in the anterior cortical regions (yellow color), and severe cerebral amyloid angiopathy (CAA) in the frontal lobe (red color, borderline association). Individual level presentation, one bar represents one subject. 6 ª 07 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association.

10 M. Tanskanen et al. Pathological Correlates Of Dementia Acknowledgments We thank prof. emeritus Raimo Sulkava for the establishment of the Vantaa 85+ Study. Tuija J arvinen and Merja Haukka are thanked for skilful technical assistance and prof. Hannu Kalimo for neuropathological advice. The study was supported by Finska l akares allskapet, Liv och h alsa Foundation, P aivikki and Sakari Sohlberg Foundation, Academy of Finland and Helsinki University Central Hospital competitive research fund (EVO). Author Contribution Dr. Maarit Tanskanen: Study conception and design, collection of data, preparation of manuscript, statistical analyses, obtaining funding. Dr. Mira M akel a: Collection of data, study design, preparation of manuscript. Dr. Irma-Leena Notkola: Study design, preparation of manuscript, statistical analyses. Dr. Liisa Myllykangas: Study design, collection of data, preparation of manuscript, funding. Dr. Sari Rastas: Collection of data, preparation of manuscript, study design. Dr. Minna Oinas: Collection of data, preparation of manuscript, study design. Dr. Perttu J Lindsberg: Study conception and design, preparation of manuscript. Dr. Tuomo Polvikoski: Collection of data, study design, preparation of manuscript. Dr. Pentti J Tienari: Study conception and design, preparation of manuscript. Dr. Anders Paetau: Collection of data, preparation of manuscript, obtaining funding, study design. Conflict of Interest Dr. Maarit Tanskanen reports no disclosures. Dr. Mira M akel a reports no disclosures. Dr. Irma_Leena Notkola reports no disclosures. Dr. Liisa Myllykangas reports no disclosures. Dr. Sari Rastas reports no disclosures. Dr. Minna Oinas reports no disclosures. Dr. Perttu J Lindsberg reports no disclosures. Dr. Tuomo Polvikoski reports no disclosures. Dr. Pentti J Tienari reports no disclosures. Dr. Anders Paetau reports no disclosures. References. Wimo A, Jonsson L, Bond J, et al. Alzheimer Disease International. The worldwide economic impact of dementia 00. Alzheimers Dement 0;9: e.. Thies W, Bleiler L; Alzheimer s Association. 0 Alzheimer s disease facts and figures. Alzheimers Dement 0;9: Plassman BL, Langa KM, McCammon RJ, et al. Incidence of dementia and cognitive impairment, not dementia in the United States. Ann Neurol 0;70:8 6.. Snowdon DA, Greiner LH, Mortimer JA, et al. Brain infarction and the clinical expression of alzheimer disease. the nun study. JAMA 997;77: Massoud F, Devi G, Stern Y, et al. A clinicopathological comparison of community-based and clinic-based cohorts of patients with dementia. Arch Neurol 999;56: Wang BW, Lu E, Mackenzie IR, et al. Multiple pathologies are common in Alzheimer patients in clinical trials. Can J Neurol Sci 0;9: Langa KM, Foster NL, Larson EB. Mixed dementia: emerging concepts and therapeutic implications. JAMA 00;9: Schneider JA, Arvanitakis Z, Bang W, Bennett DA. Mixed brain pathologies account for most dementia cases in community-dwelling older persons. Neurology 007;69: Kawas CH, Kim RC, Sonnen JA, et al. Multiple pathologies are common and related to dementia in the oldest-old: the 90 + study. Neurology 05;85: James BD, Wilson RS, Boyle PA, et al. TDP- stage, mixed pathologies, and clinical Alzheimer s-type dementia. Brain 06; Sep 0 (Epub ahead of print). PMID: Rastas S, Pirttila T, Mattila K, et al. Vascular risk factors and dementia in the general population aged >85 years: prospective population-based study. Neurobiol Aging 00;: 7.. Myllykangas L, Polvikoski T, Reunanen K, et al. ApoE epsilon-haplotype modulates Alzheimer beta-amyloid deposition in the brain. Am J Med Genet 00;: Troncoso JC, Zonderman AB, Resnick SM, et al. Effect of infarcts on dementia in the baltimore longitudinal study of aging. Ann Neurol 008;6: Mirra SS, Heyman A, McKeel D, et al. The consortium to establish a registry for Alzheimer s disease (CERAD). part II. standardization of the neuropathologic assessment of Alzheimer s disease. Neurology 99;: Braak H, Braak E. Neuropathological stageing of alzheimer-related changes. Acta Neuropathol (Berl) 99;8: Polvikoski T, Sulkava R, Myllykangas L, et al. Prevalence of Alzheimer s disease in very elderly people: a prospective neuropathological study. Neurology 00;56: Myllykangas L, Polvikoski T, Sulkava R, et al. Genetic association of alpha-macroglobulin with Alzheimer s disease in a Finnish elderly population. Ann Neurol 999;6: McKeith IG. Consensus guidelines for the clinical and pathologic diagnosis of dementia with lewy bodies (DLB): report of the consortium on DLB international workshop. J Alzheimers Dis 006;9:7. 9. Oinas M, Polvikoski T, Sulkava R, et al. Neuropathologic findings of dementia with Lewy bodies (DLB) in a ª 07 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. 6

11 Pathological Correlates Of Dementia M. Tanskanen et al. population-based vantaa 85 + study. J Alzheimers Dis 009;8: Tanskanen M, Makela M, Myllykangas L, et al. Prevalence and severity of cerebral amyloid angiopathy - a population-based study on very elderly Finns (Vantaa 85 + ). Neuropathol Appl Neurobiol 0;8:9 6.. McKhann GM, Knopman DS, Chertkow H, et al. The diagnosis of dementia due to alzheimer s disease: recommendations from the National Institute on Aging- Alzheimer s association workgroups on diagnostic guidelines for Alzheimer s disease. Alzheimers Dement 0;7: Montine TJ, Phelps CH, Beach TG, et al. National institute on Aging-Alzheimer s association guidelines for the neuropathologic assessment of Alzheimer s disease: a practical approach. Acta Neuropathol 0;:.. Braak H, Alafuzoff I, Arzberger T, et al. Staging of Alzheimer disease-associated neurofibrillary pathology using paraffin sections and immunocytochemistry. Acta Neuropathol 006;: Carlson JO, Gatz M, Pedersen NL, et al. Antemortem prediction of Braak stage. J Neuropathol Exp Neurol 05;7: Launer LJ, Petrovitch H, Ross GW, et al. AD brain pathology: vascular origins? results from the HAAS autopsy study. Neurobiol Aging 008;9: Neuropathology Group. Medical Research Council Cognitive Function and Aging Study. Pathological correlates of late-onset dementia in a multicentre, community-based population in England and Wales. Neuropathology Group of the Medical Research Council Cognitive Function and Ageing Study (MRC CFAS). Lancet 00;57: Sonnen JA, Larson EB, Crane PK, et al. Pathological correlates of dementia in a longitudinal, population-based sample of aging. Ann Neurol 007;6: Brayne C, Richardson K, Matthews FE, et al. Neuropathological correlates of dementia in over-80-yearold brain donors from the population-based Cambridge city over-75s cohort (CC75C) study. J Alzheimers Dis 009;8: Strozyk D, Dickson DW, Lipton RB, et al. Contribution of vascular pathology to the clinical expression of dementia. Neurobiol Aging 00;: Keage HA, Carare RO, Friedland RP, et al. Population studies of sporadic cerebral amyloid angiopathy and dementia: a systematic review. BMC Neurol 009;9:.. Boyle PA, Yu L, Nag S, et al. Cerebral amyloid angiopathy and cognitive outcomes in community-based older persons. Neurology 05;85: Poels MM, Ikram MA, van der Lugt A, et al. Incidence of cerebral microbleeds in the general population: the Rotterdam scan study. Stroke 0;: De Reuck J, Deramecourt V, Cordonnier C, et al. Prevalence of small cerebral bleeds in patients with a neurodegenerative dementia: a neuropathological study. J Neurol Sci 0;00: Vinters HV, Wang ZZ, Secor DL. Brain parenchymal and microvascular amyloid in Alzheimer s disease. Brain Pathol 996;6: White L. Brain lesions at autopsy in older Japanese- American men as related to cognitive impairment and dementia in the final years of life: a summary report from the Honolulu-Asia aging study. J Alzheimers Dis 009;8: Arvanitakis Z, Leurgans SE, Wang Z, et al. Cerebral amyloid angiopathy pathology and cognitive domains in older persons. Ann Neurol 0;69: Kovacs GG, Milenkovic I, Wohrer A, et al. Non-Alzheimer neurodegenerative pathologies and their combinations are more frequent than commonly believed in the elderly brain: a community-based autopsy series. Acta Neuropathol 0;6: Arvanitakis Z, Capuano AW, Leurgans SE, et al. Relation of cerebral vessel disease to alzheimer s disease dementia and cognitive function in elderly people: a cross-sectional study. Lancet Neurol 06;5: Wakisaka Y, Furuta A, Tanizaki Y, et al. Age-associated prevalence and risk factors of Lewy body pathology in a general population: the Hisayama study. Acta Neuropathol 00;06: Jellinger KA, Attems J. Challenges of multimorbidity of the aging brain: a critical update. J Neural Transm 05;: Di Marco LY, Marzo A, Munoz-Ruiz M, et al. Modifiable lifestyle factors in dementia: a systematic review of longitudinal observational cohort studies. J Alzheimers Dis 0;:9 5.. Nelson PT, Schmitt FA, Lin Y, et al. Hippocampal sclerosis in advanced age: clinical and pathological features. Brain 0;: Crary JF, Trojanowski JQ, Schneider JA, et al. Primary age-related tauopathy (PART): a common pathology associated with human aging. Acta Neuropathol 0;8: Supporting Information Additional Supporting Information may be found online in the supporting information tab for this article: Table S. Demonstrates the demographic features of the whole Vantaa 85 + population compared with the neuropathologically examined (autopsied) subpopulation. Table S. Demonstrates the results of factor analysis investigating the variance in eight quantitative variables: the neurofibrillary tangles, senile plaques, Lewy-body- 6 ª 07 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association.

12 M. Tanskanen et al. Pathological Correlates Of Dementia related pathology, infarcts (macroscopic hemispheric and deep subcortical, microscopic, MI), cerebral amyloid angiopathy (CAA), and microscopic hemorrhages (MH). Table S. Shows the distribution of combinations of the four pathologies which showed the strongest association with dementia (Braak stages V VI, presence of neocortical Lewy bodies, presence of cortical infarcts 5 mm in the anterior parts of the brains and presence of severe CAA in the frontal lobe) in the demented and nondemented subjects in the Vantaa 85 + Study population. ª 07 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. 65